Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-soc, 2009-07-01

| Channels | #perl6-soc index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
10:44 masak joined #perl6-soc
18:46 masak joined #perl6-soc
18:47 pmichaud joined #perl6-soc
18:47 ruoso joined #perl6-soc
18:51 pmurias joined #perl6-soc
19:00 masak o/
19:00 ruoso \o
19:01 masak two mentors here -- do we have any students around?
19:03 literal o/
19:03 masak literal: oh hai.
19:03 masak why don't you start?
19:04 masak (1) done since last time, (2) plan to do, (3) anything blocking?
19:06 literal ok
19:08 literal I've been working on grok. It can now output xhtml output, detect whether the source file is Pod 5 or Pod 6, and some small things like printing the name of the target file, as well as printing an index of known files. Also fixed some issues on Windows. I also made the API a bit saner and documented so it will be easier to use grok from other apps like Padre.
19:08 literal http://cpansearch.perl.org/s​rc/HINRIK/grok-0.09/Changes (see 0.06 through 0.09)
19:10 literal only blockers are the STD-like syntax lookup and the lack of a Perl 6 module filesystem structure (unless there's one I don't know about)
19:10 literal the former because I don't really know how STD works :)
19:10 masak ...and it's not really clear what 'STD-like' yet :)
19:11 literal yeah
19:11 masak s/yet/is yet/
19:11 literal as for plans, I've got a few things on the TODO list: http://github.com/hinrik/grok/blob/master/TODO
19:12 masak any preferred order among those?
19:12 literal depends
19:13 literal is there a module dir structure that people are using? where does proto install things, for example?
19:13 masak as for proto, lib/Ecosystem.pm was written with the use you mention in mind: being called by other projects for listing all Perl 6 projects, or all installed Perl 6 projects.
19:13 masak literal: proto defaults to installing 'near' itself: one floor up and then in separate directories.
19:13 masak but it's just a default.
19:13 literal ok
19:15 masak I think it all sounds promising.
19:16 literal I should start to gather some outside docs to bundle with it
19:16 literal something other than the synopses, that is
19:16 masak I'll try until next week to prime the scene for work with the syntax lookup. that's what u4x is (mostly) all about.
19:16 literal ok
19:16 masak I still think + is a good example. it occurs in a few places along the language.
19:17 masak and it's not complicated, but even (as discussed today) numification of lists has pitfalls worth mentioning.
19:19 literal one other thing... I understand it was you who originally wrote what ended up as the S32 function lookup code in Padre::Plugin::Perl6 ?
19:19 masak uh, yes.
19:19 literal that seems like something that should be in grok, like perldoc -f
19:19 masak definitely, but without the -f.
19:19 literal I might work with azawawi on porting that
19:20 literal yeah
19:20 masak we might start with taking the docs from S32, but ultimately u4x will have its own docs on that too.
19:20 masak see u4x/TODO.
19:21 literal ok
19:22 masak pmurias: you there?
19:22 pmurias yes
19:22 masak do you have something you'd like to report?
19:22 pmurias writing a commit message atm, please wait a bit ;)
19:22 ruoso I'm missing your blog posts, btw
19:26 pmurias just commited a simple implementation of signature sorting
19:26 pmurias ruoso: i'll blog about it then
19:28 masak hm, I feel like we lost momentum a little.
19:28 masak pmurias: you didn't go away to blog now, did you? :)
19:29 pmurias not yet...
19:29 masak pmurias: ok, anything else to report?
19:29 masak (1) done since last time, (2) plan to do, (3) anything blocking?
19:29 pmurias got hypnotised by watching the tests slowly pass....
19:30 ruoso yeah... mildew is starting to be strangely slow
19:30 masak the curse of Perl 6 implementations, it seems.
19:31 * pmurias notices rakudo is getting faster and faster...
19:31 ruoso but we intentionally didn't make any optimization at all
19:31 ruoso and we know there are a lot that can be done in that area
19:32 pmurias Multi methods are used in some places already, and they are getting really slow now that their implementation is growing
19:32 masak ruoso: that really sounds like Rakudo a while ago. :)
19:33 ruoso simple things like, punned classes are not cached would buy us a lot of time
19:34 ruoso pmurias, do you think the speed will start to be a blocker soon?
19:35 pmurias the use of "make test" starts to require strategic timing
19:36 pmurias i don't think the multis are fast enough to build anything on them
19:37 ruoso pmurias, I have a feeling that getting mildew to emit LOST code instead of mold would allow us a lot of tricks
19:38 pmurias going towards compilation seems the reasonable choice
19:39 pmurias ruoso: do i need to implement is rw and is copy?
19:39 ruoso I think we can go a long way without it
19:40 ruoso pmurias, the most promising thing about going to lost, is that we can do the test for native types and access the c functions directly instead of doing another method call
19:41 ruoso that alone should avoid *LOTS* of indirections
19:43 ruoso pmurias, do you think it would be too hard to change the compiler to generate lost frames?
19:44 pmurias m0ld is pretty flexible
19:44 pmurias the harder part would be getting all that compiling down known RI's down
19:45 pmurias i'm unsure if we want to have it in Perl 5/Perl 6 or haskell
19:47 pmurias s:1st/down//
19:47 ruoso the haskell dependency is confusing some people
19:51 pmurias we could have the mold to LOST compiler in Perl 6
19:51 ruoso that'd be cool
19:54 literal LOST code?
19:54 * masak feels LOST
19:54 pmurias why are they called LOST frames by the way?
19:54 ruoso lack of creativity
19:55 masak but what does it stand for?
19:55 ruoso LOwlevel STate machine ;)
19:56 masak :)
19:57 pmurias what i feel i must fix first before moving on to optimalisation is native stuff not implementing the repr api and not accepting containers
19:58 ruoso pmurias, that seems like a good choice
20:01 masak anything more we should discuss tonight?
20:03 masak ok, let's wrap it up, then.
20:03 masak thanks everybody, and see you here next week!
20:03 pmurias see you
20:04 masak left #perl6-soc
20:10 pmurias left #perl6-soc
20:49 pmichaud left #perl6-soc
21:01 ruoso left #perl6-soc

| Channels | #perl6-soc index | Today | | Search | Google Search | Plain-Text | summary